Tyndall-John: Arbiter (Step 3)

Project: John Tyndall's radiant-heat, gas-absorption and atmospheric-vapour work (1859-1892). Inputs: facts.json, claims.md, physicist.md, mathematician.md, contrarian.md, deep/, sites.md, research/. Date: 2026-10-08. Tags: [read] from the project documents (as quoted in the reviews and deep reads), [web] with URL, [memory] with confidence, [calc] my own arithmetic (numpy in the project venv).

Source gap and source limits (repeated from claims.md)

Not a full source gap: the five key works are readable (OCR layer, numbers checked on page images by the deep readers). The limits that carry into every verdict: - Everything in the folder is by Tyndall, an interested party on his own priority (calorescence, water-vapour absorption, the glacier theory, Mayer). No independent replication or critique is in the folder; all of it comes from the Investigator's web notes, which are second-hand to the reviewers (research/files/ is empty, nothing was downloaded). - Missing from the folder and unread by anyone: journal p.1 and Plate I of the 1866 paper (Muller's solar curves), the unpublished galvanometer calibration, Magnus 1861/62, Lecher and Pernter 1880/81, Strachey July 1866, Foote 1856 text, Tyndall's Contributions to Molecular Physics (1873), Mayer's own paper. - (Revised, see QE revision) Heat Ch. XI (pp.362-378), XII, XIII and XIV (647-684) have now been deep-read with page images; the CO2 pressure numbers (48.0 to 74.3 per cent) are image-confirmed. Other chapters of Heat remain digest only. No verdict below turns on them alone. - Where a conclusion rests on [memory] alone it is marked.

Recomputed numbers where the reviewers differ

1. Blackbody infrared:visible ratio (claim 3). Physicist, mathematician and contrarian quote slightly different numbers (3000 K: 7.7, 7.8, 6.9; sun: 1.07, 1.07, 0.93). The difference is only the band definition. [calc] Infrared 0.76-15 micrometres to visible 0.4-0.76: 18.3 (2500 K), 10.5 (2800 K), 7.74 (3000 K), 4.26 (3500 K), 3.21 (3800 K), 2.73 (4000 K), 1.07 (5772 K). Infrared above 0.78 to visible 0.38-0.78: 15.9, 9.3, 6.9, 3.8, 2.9, 2.45, 0.94. The arc's 7.7 corresponds to about 3000 K in the first convention (3004 K; 8 gives 2977 K) and about 2900 K in the second. All three reviewers are right; the physics is unchanged. The sun: about 1.0 at the top of the atmosphere in either convention, so Muller's "about 2" is high by roughly a factor two (cause unknown; his curves are not in the folder).
2. Mayer-type arithmetic (claim 6). [calc] With air heat capacity 0.24 and density 0.08073 lb/ft3: pressure 14.696 psi, gamma 1.40 gives 777.9 (matches the modern 778); gamma 1.421 gives 750.1; with 15 psi and 1.421 about 766 before Tyndall's other roundings (1.29 oz, 490 F), which bring it to his printed 771.4. So the mathematician's "two compensating errors" holds: rounding 14.7 up to 15 psi raises J by 2.1 per cent, and using gamma 1.421 instead of 1.40 lowers it by 3.6 per cent. Two corrections to the reviews: (a) the mathematician's sensitivity formula d ln J / d ln gamma = 1.67 is wrong; the correct value is 1/(gamma-1) = 2.5, which agrees with his own table (797 at 1.40 to 770 at 1.42 is -3.4 per cent for +1.4 per cent). The "gamma lowers J by about 2.5 per cent" he gives is too small for the 1.5 per cent gamma difference; 3.6 per cent is right. (b) The contrarian's "a 1 per cent change in the ratio changes 2.8 by about 10 per cent" is wrong; [calc] 1 per cent in gamma changes the 2.8 by about 2.5 per cent. The real point stands: the result is a difference of two near-equal heats and is sensitive, and the agreement with Joule's 772 is partly luck.
3. Mayer's own figure. The contrarian says Mayer's 1842 figure is 365 kgf-m/kcal (about 665 ft-lb per lb per F, [calc] 365 x 3.2808 / 1.8 = 665); the mathematician could not confirm the 1842 link. [web] https://en.wikipedia.org/wiki/Julius_von_Mayer says the value "at first as 365 kgf-m/kcal" comes from Mayer's 1845 booklet, and does not give a number for the 1842 paper. So: 365 is attested for 1845; its attribution to 1842 is unconfirmed. Either way 771.4 is Tyndall's reconstruction with his own data, not a figure of Mayer's.
4. Ten feet of air (claim 2). Tyndall's own 7.9 per cent in the open tube extrapolates, by simple exponential scaling, to 37.7 per cent if the 7.9 applied to 21 inches of path and 17.7 per cent if to 51 inches [calc]; neither is what he claimed, and the path length is not stated in the text. Planck fractions of a 288 K body: 7.5 per cent at 5.5-7.5 micrometres, 38 per cent beyond 17, 19 per cent at 13-17, 25 per cent in 8-12 [calc]. The physicist's and the contrarian's order-of-magnitude estimates ("a few per cent up to about 10 per cent" for 3 m of humid air) are crude band counts, the mathematician's "a few per cent" is a low-confidence memory of Hottel charts. I tried to check the Hottel emissivity chart on the web; the search returned the chart's existence (e.g. https://thermalradiation.net/On-line appxs text 9-27-2024.pdf, Alberti charts) but no readable value, so the check failed. This number stays unresolved; a line-by-line calculation would settle it (see "What would settle it"). (withdrawn, see QE revision item 3)
5. Water-vapour figures (claim 2). Physicist is right that claims.md garbled "70-90 times per unit amount of the dry air": 70-90 is the absorption by the vapour actually present (about 0.45 per cent of the molecules) relative to the whole air; per molecule it is 80 x 200 = 16,000 (or 80 x 222 = 17,800 with his 0.45 per cent). The two figures differ by about 200 and must not be merged.
6. Units of "times air" (claim 1). Mathematician is right and the physicist agrees: the 1-inch table compares each gas with air at the same pressure; the vapour and olefiant-gas pressure tables compare with a whole atmosphere of air. [calc] CO2 972 per 1 inch = 32.4 whole-atmosphere-air units; ammonia 5460 = 182; olefiant gas 6030 = 201. The ratio 30 inch / 1 inch absorption is only 2.8 (CO2) to 6.6 (NH3): saturation.
7. "One-ninth" versus 11:89. Physicist's note stands: 11:89 is 1:8.1, one-ninth of the dark heat is 10 per cent of the total; the 1865 and 1866 figures are different measurements, a 10 per cent discrepancy. claims.md was wrong to call them the same.

Corrected claim list (Step 1 statements the reviews corrected)

• Claim 1 (QE revision): reciprocity rests on 4 static and 13 dynamic measured pairs; rank order agrees, proportionality only roughly (absorption/radiation 0.40-0.87). See QE revision.
• Claim 1: "air = 1" is a floor, not zero; all the large ratios are lower bounds, specific to his source, tube and calibration. Add chlorine 60 and bromine 160 as anomalous (homonuclear diatomics should be infrared-inactive; probable traces of impurity or the floor effect, [memory: high] on the selection rule, cause an inference), and ozone, an elemental absorber Tyndall himself found.
• Claim 2: see item 5 above. The "ten feet" statements are two different ones: Rede lecture "10 to 15 per cent for a ten-foot column of ordinary air" (no experiment cited) and Heat para 480 "at least 10 per cent of terrestrial heat within ten feet" (extrapolated from a 4 ft tube and a 100 C source).
• Claim 3: "sun about 2" is Muller's figure quoted by Tyndall, not Tyndall's result.
• Claim 4: "reverse of Stokes's rule" is Tyndall's description; Stokes himself wrote on 13 Jan 1865 that the process "is not a little different from fluorescence" and that the name was not quite correct [web: https://epsilon.ac.uk/view/tyndall/letters/Tyndall2215, per Investigator]. claims.md should carry this beside the claim.
• Claim 5: Tyndall's sentence that red waves are "immensely more powerful than the waves of blue" is wrong if energy per wave is meant (photon energy is proportional to frequency; a 420 nm photon carries 700/420 = 1.67 times the energy of a 700 nm photon [calc]). The experiment (visible light decomposes nitrite of amyl, ultra-red does not) is consistent with the photon picture.
• Claim 6: 771.4 is Tyndall's reconstruction (items 2, 3). Rumford's 107 is a temperature in F after one hour; 23.5 versus 28 minutes is the goldbeater's-skin vessel result (coating, not colour), per the New Fragments deep read.
• The Rede lecture's vapour table (CS2 47 ... acetic ether 612 at "1/30 atmosphere") does not match Heat's (62 ... 1195 at 1.0 inch); ratio Heat/Rede 1.3 to 2.6; apparatus or source unstated. Do not mix the two tables.
• Strachey: Heat gives a 4-25 March 1850 series; Tyndall's 1882 Bakerian describes the same 1866 paper as Madras data of 1841-44. Both can be in one paper; unresolved (the mathematician's and contrarian's searches for the 1866 paper failed).
• Small internal slips by Tyndall, kept in the record: nitrous oxide "1800" in prose, 1860 in table; several printed "calculated" values in Heat tables 7b-7d off by 1-3 per cent; the Rede lecture says half-a-tenth inch of water stops "no single ray" of terrestrial heat, stronger than Heat's own data (opaque at 0.27 inch to a hydrogen flame; over 97 per cent cut at about 0.04 inch).

Verdicts by claim

Claim 1. Elementary gases transparent to radiant heat, compounds absorb strongly (972, 5460, 6030 times air); radiating power equals absorbing power

Agreement: all three: Established. Disagreements: none on the verdict, only on emphasis. Physicist and mathematician stress the numbers are lower bounds; the contrarian adds that Franz found 3.54 per cent apparent absorption in dry air in his tube (the same baseline problem) [web: Very 1900, https://archive.org/details/atmosphericradia00veryrich, per Investigator]. Stronger argument: all three, because the qualitative pattern is independently confirmed: Lecher and Pernter on dry gases (1880), Rontgen with a different method (1881/84), photoacoustic work (Bell, Tyndall 1880-81) and a century of spectroscopy [web, per Investigator]. Priority: Franz 1855 and Foote 1856 reported gas absorption earlier; Jackson 2019 treats it as simultaneous discovery and notes Tyndall did not cite Foote [web: https://royalsocietypublishing.org/doi/10.1098/rsnr.2018.0066, per Investigator]. This is Tyndall's priority being qualified by a historian, not a physics question. Macdonald's viXra paper (2015, unrefereed, the author's claim) that N2 and O2 absorb because they are Raman-active confuses polarisability change (Raman) with dipole change (infrared) [web: https://vixra.org/abs/1504.0165, per Investigator; memory: high on the selection rules]; weak collision-induced absorption exists [memory: medium] and is orders of magnitude below CO2. Verdict: Established (qualitative pattern, ordering, Kirchhoff reciprocity). The numerical table: Established only as a source-specific ranking; magnitudes are lower bounds, not constants. Settles the rest: the same table measured with a modern FTIR on the same source spectrum (see below).

Claim 2. Water vapour controls Earth's heat loss ("blanket"); at least 10 per cent of terrestrial heat intercepted within ten feet; CO2 minor

Split into parts; the three reviewers agree on the first, third and fourth and differ on the label for the second and on how to label the ratios. - 2A/2C/2D, vapour absorbs strongly, is the dominant absorber of the lower atmosphere, controls night cooling, deserts, mountain ranges; transparent to sunlight, opaque to earth heat. Agreement: Established. The dispute was real and long: Magnus (1861/62) and Buff found no vapour action and blamed films on the rock salt; Lecher and Pernter (1880/81) found "moist air a little more transparent than a vacuum"; Very 1900 defended Tyndall but asked for "further elucidation" of the humid-air table; Rontgen 1884 found moist air does absorb by an independent method and said many physicists still thought the question open [web, all per Investigator]. Tyndall answered with controls the reviewers judge careful (rock-salt plates removed, open tube, no tube, air from seven sites, a table of proportionality constant to about 2 per cent over a six-fold pressure range [calc: observed/pressure 3.2-3.28 per inch]). A result "more transparent than a vacuum" is itself a sign of a systematic error in that apparatus (physicist, contrarian; an inference, no source). The record does not read as suppression: the dispute was conducted in print and by letter, Tyndall's reply was printed, and it was resolved by method change and spectroscopy. One caution I add from the sources: much of the correspondence is Tyndall's side (Hooker: "Magnus denies the action; but he has not a foot to stand upon"; Airy and Herschel approval reported via Tyndall), so the hostile side is less documented in the folder than his; Magnus's and Lecher-Pernter's raw data were not seen by anyone. Added caveats (kept): the Strachey series is one chosen "representative" series (r = -0.98, slope -22.5 F per inch of vapour tension, n = 9 [calc, mathematician]) with uncontrolled wind, cloud, ground and a dew-point floor (night temperature cannot fall far below the dew point because condensation releases latent heat), so it does not isolate the radiative mechanism; it is nevertheless in the right direction and consistent with modern physics (about 33 K between a 255 K blackbody Earth and the 288 K surface [calc, physicist]). - 2B "at least 10 per cent within ten feet". Reviewers labelled it Plausible untested (physicist, contrarian) or Never properly tested (mathematician). Argument: it is a testable number that was never tested on terrestrial radiation, so "never properly tested" is literally true; but it is consistent in order of magnitude with humid air and a broad source, and "tested" would only need a calculation, not an experiment. Estimates on the table range from a few per cent to about 10 per cent for typical surface air; my own web check failed (item 4). I assign Plausible, untested as stated: a laboratory extrapolation, at the high end of what the reviewers can justify for typical air, and "certain" (Tyndall's word in the Rede lecture) too strong. It was not dismissed without testing; nobody tested it. - 70-90 times and 16,000 per molecule: Not testable as stated as constants (lower bounds against a purity-limited reference); 16,000 is a product of two rounded numbers; order of magnitude only. - 2E CO2 "minor": Historically superseded but instructive. The attribution to Tyndall of "CO2 is unimportant" is a silence turned into a ranking: he never apportioned effects, and his own tables show CO2 as a strong absorber per molecule (972 at 1/30 atmosphere) and, against a carbon-monoxide-flame source, absorbing 48 to 74 per cent [read: page image p.431, QE deep read]. Current model attribution is about 50 per cent water vapour, 25 per cent clouds, 20 per cent CO2 of the present greenhouse effect, with the authors stressing the shares depend on how overlap is allocated, and that vapour responds to temperature whereas CO2 largely does not [web: https://www.giss.nasa.gov/research/briefs/2010_schmidt_05/, Schmidt et al. 2010]. The climate argument itself appears 1861-65, not in the 1859 Note [web: Jackson 2019]. Overall: Established (core: vapour absorbs and its layer controls heat loss); 10-foot number Plausible, untested as stated; ratios Not testable as stated; CO2-minor Historically superseded but instructive.

Claim 3. Invisible radiation exceeds visible (arc 7.7 by area, 8 by filter; sun about 2); iodine in carbon bisulphide passes heat, blocks light; dark rays ignite materials and make platinum and carbon glow

Agreement: Established for the arc and the filter; solar "about 2" disputed only in label (physicist: high; mathematician: Never properly tested; contrarian: Never properly tested here). I side with the mathematician and contrarian on the label: Muller's data are not in the folder, so it was not tested here, and by [calc] it is high by about a factor two against modern values (infrared:visible about 1.0 for a 5772 K body; measured top-of-atmosphere solar spectrum roughly 49 per cent infrared and 43 per cent visible, [memory: high]; ground level lower still). Strength: 7.7 and 8 agree within 4 per cent by two methods that share only the eye's red limit; 7.7 corresponds to about 3000 K, plausible for glowing carbon points of an arc (the crater is nearer 3800-4000 K [memory: medium-high], so the lamp's effective ratio is a mixture). Weakness: both depend on a calibration table the reader cannot see; raw readings give 94.3 per cent transmission for the iodine cell against Tyndall's 86.8 per cent after calibration, and 98.6 against 94.8 for the 2-inch bisulphide cell [calc, mathematician]. The assumption that all absorption by iodine is visible light is stated by Tyndall as an inference ("may be inferred") and he concedes some extra-red absorption. The physicist and mathematician state the direction of the resulting bias differently (the mathematician's wording is itself tangled); neither shows it alters the conclusion, and I do not resolve it. Akin and Griffith independently got platinum incandescent in sunshine (1863/65) [web, per Investigator]. No independent measurement of the arc ratio was found. Safety note carried from the physicist: Tyndall's undefended-eye test is dangerous; "retina not sensibly affected" does not mean safe. Verdict: Established (arc ratio and filter, source-specific). Solar "about 2": Never properly tested (cited from Muller, apparently high).

Claim 4. Calorescence: transmutation of heat rays to higher refrangibility, the reverse of Stokes's rule; blue glasses give pink glow (gap)

Split, all three reviewers agree on all three parts. - 4A observation (dark focused rays make platinum and carbon glow, also in vacuo, continuous spectrum): Established. Ordinary thermal incandescence; the absorbed power heats the foil, which then emits a Planck spectrum. Draper point about 798 K [web: https://en.wikipedia.org/wiki/Draper_point, per physicist]; a thin foil radiating from both faces needs about 4.6 W/cm2 for 798 K, about 57 W/cm2 near 1500 K [calc, physicist: 2 x 5.67e-8 x 798^4 = 46 kW/m2], easily reached by a lamp of about a kilowatt with a reflector. - 4B interpretation as ray-by-ray transmutation: Historically superseded but instructive. Tyndall's paper holds two pictures (thermal emission, p.3 and 7; rays returned "at a quicker rate", p.17-18). His private words to Croll, 14 Jan 1865 ("you hit the mass with waves of slow recurrence, and ... obtain, in return from the mass, waves of quick recurrence", [web: https://epsilon.ac.uk/view/tyndall/letters/Tyndall2216]) already describe heating. Stokes's reply (above) is in the record. A single infrared quantum cannot become one visible quantum; energy is conserved, the second law is satisfied because the source (about 3000 K) is hotter than the glowing foil (about 1500 K) [physicist]. Priority dispute with Akin: unresolved by the folder; every account of it in the folder is Tyndall's (an interested party) plus the Epsilon letters; Sabine and Armstrong (British Association) sided with Tyndall, Tait (1868) credits Akin with "Calcescence or Calorescence" [web, per Investigator]. - 4C pink glow, "gap in the calorescence": Never properly tested. Tyndall wrote it as conditional and asked for confirmation; no one confirmed or refuted it; it was ignored, not dismissed. Two reviewers independently note his own printed glass table (blue glasses pass blue, a green band and extreme red) gives a pink tint of a continuous glow with no gap, and dim-light colour judgement is poor; this is their inference, untested. Verdict: observation Established; transmutation/reverse-Stokes Historically superseded but instructive; pink gap Never properly tested.

Claim 5. Absorption is molecular, selected by resonance; conserved when compressed or condensed (conservation of molecular action, thermal continuity)

Split. Reviewers differ in labels: physicist Plausible untested for the density/phase invariance; mathematician Established for it; contrarian Plausible untested for the weak form and Tested, not supported for the strong form. Arbitration: Tyndall's data are tight (18.4 versus 18.8, 25.7 versus 25.6, 23.5 versus 23.4, 32.1 versus 32.6; liquid ether 67.2 versus vapour 66.7 per cent; molecule counts 9.62e-4 versus 9.66e-4 mol/cm2 [calc, matches]) but they are single tabulated pairs, no repeats or errors, one source type (broadband), vapours without hydrogen bonding, density range of a factor 3.5. Independent support for molecular origin exists (Rontgen, photoacoustics). Independent test of the exact pairs: none on record. Modern spectroscopy shows pressure broadening and phase shifts change band shapes (liquid water differs strongly from vapour; liquid water absorbs across 8-12 micrometres where the vapour is nearly transparent) [memory: high, not web-checked]. I therefore split: - Molecular (not film) origin and resonance selection: Established (the quantum dipole-transition picture replaces the resonance-of-periods language). - Weak form (integrated broadband absorption per molecule roughly independent of density and phase, for these test vapours): Plausible, untested independently (supported by Tyndall's own data to about 0.5 percentage points). I do not use "Established" because the supporting tests are all his. - Strong form (independent of density and phase generally, including water): Tested, not supported, on [memory: high] spectroscopy alone; no source fetched. Note that it was never Tyndall's tested claim in this form. (superseded by QE revision item 11) - "Ether entangled in the molecule" and the periods-of-atoms mechanism: Historically superseded but instructive. - The sentence about red waves being "immensely more powerful" than blue: Internally inconsistent if read as energy per wave (mathematician); an error of statement, not of the experiment.

Claim 6. Radiant heat and light are the same wave motion; heat is a mode of motion (Rumford, Davy, mechanical equivalent 772; Mayer reconstructed 771.4)

Agreement: Established. Identity of heat and light as electromagnetic waves (Maxwell 1865, Hertz 1887-88, [memory: high]); heat as molecular motion; mechanical equivalent within 1 per cent (modern 778 ft-lb per Btu, [memory: high]). The ether as a physical carrier: Historically superseded but instructive. Qualifier added by the mathematician and confirmed by my recomputation (item 2): the 771.4 agreement with Joule's 772 is partly accidental (two compensating errors), so it is not evidence of the method's accuracy; and it is Tyndall's reconstruction, not Mayer's number (items 3). Tyndall's single-source claim that he magnetised obscure heat on a larger scale (about "150 of the lower degrees of the galvanometer") than De la Provostaye and Desains (two or three degrees) is quoted and weighed, not dismissed: it has no independent confirmation in the folder, and the deflection scale is non-linear.

Claim 7. Glacier ice is not viscous; regelation; Ice Age needs abundant vapour and strong sun (Heat Ch. VI)

Agreement: regelation Established. Strong form: physicist Tested, not supported; mathematician and contrarian Historically superseded but instructive. Arbitration: the physical test exists (Glen 1955: polycrystalline ice creeps continuously under constant stress, strain rate roughly proportional to stress cubed, [web: https://data-ww3.ifremer.fr/BIB/Glen_PRSLA1955.pdf, per physicist]) and it contradicts "not viscous", so "Tested, not supported" is the scale term that applies; but it was a reasonable position in 1857-68 and Tyndall himself conceded ice "behaves like a viscous body" under pressure and that viscosity "accounts, certainly, for half the facts" [read: Heat p.211-212]. Forbes called the views "complementary" [web: https://www.sciencehistory.org/stories/magazine/under-pressure/, per Investigator]. His "most delicate experiments" that failed to detect stretching are cited to no one, with no numbers. Regelation survives in basal sliding and crevasse healing [memory: medium; Weertman 1957 not checked]. Pressure melting alone is tiny: -0.0075 K per atmosphere [calc, mathematician]. Verdict: regelation Established; "ice is not viscous, flow is fracture plus regelation" Tested, not supported (strong form), Historically superseded but instructive in context; Ice Age energy argument (5 lb of iron per lb of vapour, 1000 x 10 / 2000 [calc]): arithmetic internally consistent, mechanism Not testable as stated (the "improved condenser" is never specified), and superseded by orbital forcing with feedbacks [memory: high on Milankovitch; not web-checked]. Tyndall's own words: "I have no new hypothesis to offer". Observation recorded: this is the one claim where the interested party won the argument in his lifetime and the later evidence went the other way.

Verdict table

What would settle it

• Claims 1, 5 (numbers, phase/pressure invariance): measure the Tyndall gas table with a modern FTIR spectrometer or a thermopile and rock-salt tube (the Royal Institution holds a 698 mm x 85 mm tube dated 1861 [web: https://www.rigb.org/explore-science/explore/collection/john-tyndalls-radiant-heat-apparatus, per Investigator]) using the 1-inch pressure, copper-plate source and highly purified air, and the liquid versus vapour ether pair at equal molecule count. A line-by-line (HITRAN) calculation of the same cases is the cheap version.
• Claim 2B (ten feet): a line-by-line radiative-transfer calculation of the transmittance of a 3 m path of humid surface air (water plus 400 ppm CO2) against a 288 K blackbody, plus the same for a 373 K source and a 4 ft tube to check the 7.9 per cent. Also the original path length of the open-tube experiment (Tyndall's Contributions to Molecular Physics, 1873).
• Claim 2 (Magnus, Lecher-Pernter): the raw data of Magnus 1861/62 and Lecher and Pernter 1880/81, and Strachey's July 1866 paper (which settles March 1850 versus Madras 1841-44).
• Claim 3: Muller's solar curves and the 1866 paper's journal p.1 and Plate I (jstor-108936); the galvanometer calibration table; a modern spectral radiometer measurement of a carbon arc and a 3000 K lamp, with and without an iodine-bisulphide cell (do this with eye protection).
• Claim 4C (pink glow): view a platinum foil heated to dull red, faint, through a blue glass whose transmission has been measured with a spectrometer; a pink tint from a continuous spectrum with the glass's measured bands would end the "gap" question. Claim 4 priority: Akin's BA Report 1863 and the nine letters of the 1864-65 correspondence (Tyndall Correspondence vols 8-9).
• Claim 6: arithmetic only; [calc] the Mayer-style chain with exact pressure, density, heat capacity and gamma gives 778.
• Claim 7: already settled by creep experiments (Glen); remaining historical question is who performed the "most delicate experiments" on stretching ice (Tyndall's Glaciers of the Alps).

Open questions

• Which path length (21 or 51 inches) applied to Tyndall's 7.9 per cent, and what is the right ten-foot figure for real air.
• What Muller's solar ratio of about 2 was measuring (instrument, prism, thermopile response).
• Whether the Strachey series of Heat is the March 1850 data or Madras 1841-44, and for what place.
• Why Heat's vapour table differs from the Rede lecture's by a factor 1.3-2.6 (different apparatus or source; unstated).
• Why chlorine 60 and bromine 160 appear as absorbers (impurity, floor effect, or an unrecognised effect) [physicist's inference only].
• The Tyndall-Akin priority: who published what first, given Tyndall's "nearly nineteen months" held-back experiment (Epsilon 2167, per the contrarian).
• Whether the lower precision of the Magnus-era experiments or something else explains their negative results; the reviews give plausible artefacts (heated glass plates as secondary radiators, vapour films) from Very 1900 and Tyndall, each partly an interested party's account.

Absence-of-evidence analysis (where the record goes dark)

1. Akin after 1865 (Károly Akin, born Kohn, 1830/32-1893). After the calcescence dispute his physics publications stop; he became a Hungarian academic and political figure (Academy member 1868) and died in poverty (Nevpont, per Investigator). The record is silent on physics where it might have continued. Competing explanations: (a) the dispute and its outcome drove him out (Sabine, Armstrong and the Royal Society establishment sided with Tyndall; Armstrong wrote the BA "could not ... confer upon Dr. Akin any exclusive privileges", [web: Epsilon letters 2075-2208, per Investigator]); for: the timing; against: nothing in the folder says he was blocked from publishing, and he kept an academic career. (b) a change of interests and career in Hungary; for: his Hungarian posts; against: none found, the biography is short. (c) no laboratory post or funding in Britain; for: he worked with Griffith at Oxford with a borrowed 2-ft mirror; against: no source states it. (d) deliberate suppression: no evidence; I cannot rule it out, because every account of the dispute in the folder is Tyndall's and the lopsided contest between an outsider and the Royal Institution professor is real. What would decide: Akin's letters in the Tyndall Correspondence vols 8-9, the Hungarian Academy archive, the 1863 BA Report text.
2. The "distinguished authority", "eminent men" and "most delicate experiments" on ice stretching: Tyndall gives no names or data. The silence is in Tyndall's text, not in the world's record; the archive to look at is Tyndall's own memoirs and the Forbes papers (St Andrews). Competing readings: a real but unpublished experiment; an informal remark from Forbes's opponents; rhetorical overstatement. No evidence decides.
3. The 1859 first positive result (Tyndall's journal of 18 May 1859, RI MS JT/2/13/1158, cited but not seen) and the unpublished calibration table: documents that exist (or existed) and have not been examined.
4. Magnus's reply and the Geneva printing: per Soret (Epsilon 2040, per Investigator), Geneva Archives printed Magnus's memoirs and not Tyndall's reply. This is a single-source observation by a correspondent; it indicates uneven publication but not suppression. The main vapour dispute does not go silent: it moves through Buff, Lecher and Pernter and Rontgen to spectroscopy.
5. Mayer's 1842 calculation: Tyndall says Mayer's first paper gave no details; his own paper is not in the folder; the 1842 figure is unconfirmed (item 3).

Investigator findings kept in the record (whatever the verdict)

Foote 1856 and Franz 1855; the 1859 Note (first crude test negative, then "very large effects", no numbers, no climate); Forbes's letter of 15 July 1859 (surprise at magnitude); Tyndall to Hooker and Herschel on Magnus; Airy "we should be in a way for beginning Meteorology" (reported through Tyndall's side); Stokes on the name; Croll; Akin and Griffith and the 1865 history; Tait 1868; Babbitt 1878 using calorescence in an alternative colour theory; Macdonald (viXra 2015); Protons for Breakfast commentary (no replication); the Royal Institution tube; Science History Institute on Forbes ("utterly annihilated", "complementary", Forbes "disheartened and hopeless" 1864, per the contrarian).

Corrections to earlier documents

1. claims.md claim 2: "70-90 times per unit amount of the dry air" is wrong: 70-90 is for the vapour actually present; per molecule it is about 16,000. Evidence: Heat p.380, 387-388, 401.
2. claims.md "one-ninth ... same ratio as 11:89": wrong; 11:89 is 1:8.1.
3. claims.md "about 797 with the modern ratio of specific heats 1.40" presented as modern value: the modern value for the method is 778 once 14.696 psi is used; 797 follows from keeping Tyndall's 15 psi with 1.40.
4. mathematician.md section 6 and correction 1: sensitivity d ln J / d ln gamma = 1.67 is wrong, correct value 2.5; "gamma lowers J by about 2.5 per cent" should read 3.6 per cent.
5. contrarian.md claim 6: "a 1 per cent change in the ratio changes 2.8 by about 10 per cent" is wrong; about 2.5 per cent.
6. contrarian.md correction 1: Mayer's "1842 figure is 365" is not supported by the cited source (365 is attested for the 1845 booklet; 1842 unconfirmed).
7. mathematician.md claim 3a: the sentence about the direction of bias from extra-red absorption by iodine is self-contradictory as written; the physicist's version (a larger extra-red absorption makes the true visible share smaller and the true ratio larger) is the straightforward reading, and neither version changes the verdict.
8. claims.md claim 4: "reverse of Stokes's fluorescence rule" is Tyndall's own wording, which Stokes questioned in 1865.
9. Deep-read Heat digest: the OCR figure for the Heat 1-inch vapour table differs from the Rede lecture's; flag in the primer that the two tables are not interchangeable.
10. Heat deep read, Mayer line: "sits on a ratio that gives about 770 only because 1.42 was the accepted value" is incomplete: two compensating errors (15 psi and gamma).
11. Investigator notes (Epsilon): the editors' remark that Tyndall "did not write or publish on the subject" from Nov 1864 to 1865 conflicts with his Phil. Mag. 29 (1865) 218-31 history of calorescence (contrarian, correction 5); the month is not given. Check before citing.

Web checks done (budget 8)

1. Mayer 365 kgf-m/kcal: https://en.wikipedia.org/wiki/Julius_von_Mayer; confirms 365 from 1845 booklet; 1842 unconfirmed. Retagged [web].
2. Hottel emissivity of water vapour at 0.1 atm-ft: failed (no readable value); the ten-foot figure stays an estimate. The remaining [memory] facts verdicts rely on (infrared selection rules, Kirchhoff, Glen's law, Milankovitch, Maxwell and Hertz, 778 ft-lb per Btu) are high confidence and were checked by the reviewers where noted; the phase dependence of water's spectrum, solar infrared fraction and Weertman sliding are [memory] and not web-checked here.

QE revision (2026-10-09)

Basis: qe.md (verdict MINOR) and the new deep reads in deep/Heat-35e701.md (Ch. XI paras 435-466; Ch. XII paras 497-556; Ch. XIII paras 557-646; Ch. XIV paras 647-684; Magnus note pp.250-254 and p.533) and deep/New-fragments-John-Tyndall-626498.md (Titlis 1866). Tags as above; I did not re-derive the deep readers' page-image readings. No claim's headline verdict changes. Sub-verdicts are refined as listed.

Changes to verdicts and caveats

1. Claim 1, reciprocity (radiating power equals absorbing power): Established stands, but the quantitative word "equals" is not what the numbers show. [read: Ch. XI deep read s.5] Measured pairs: 4 static (CO 12/18, CO2 18/25, N2O 29/44, olefiant 53/61) and 13 dynamic (2 gases, 11 vapours). Rank order agrees in every one of the 55 vapour pairs; absorption/radiation runs 0.40 to 0.61 across vapours, 0.6 to 0.7 for the two gases, 0.66 to 0.87 for the static pairs. Single runs, no error bars; Tyndall's own repeats differ by up to 4 galvanometer degrees; readings of 60 to 70 degrees sit on the non-linear end of the scale. The dynamic pairs measure heating and cooling of one emitter, so they test symmetry rather than Kirchhoff independently (deep reader's inference). Verdict: reciprocity Established as rank correlation and rough proportionality; exact equality rests on theory (Kirchhoff, Balfour Stewart), not on these data.
2. Claim 1, ozone: the figures are now read: 20, 34, 47 (cells with 4, 2, 1 sq in plates), 65, 85, and 136 with the cell in ice and salt, times the carrier electrolytic oxygen [read: image p.366-367]. Ozone concentration was "perfectly unmeasurable by ordinary means", so these rank preparations; they are not an absorption coefficient. Tyndall's conclusion (ozone is oxygen atoms in oscillating groups) is a single-author inference that he says later chemistry confirmed; no source for that in the folder. The earlier text on ozone stands as correct but unquantified.
3. Claim 2B ("at least 10 per cent within ten feet"): wording corrected, verdict unchanged (Plausible, untested as stated). My item 4 above said the extrapolations were not "what he claimed". That was misleading: Tyndall wrote "at least", and every extrapolation (17.7, 37.7, about 23 per cent from his 4 ft polished-tube "over 10") exceeds the floor [calc, QE]. The real questions are saturation and source. New evidence from his own pages [read: image p.428-429]: the same 4 ft polished tube of undried laboratory air absorbed 5.8 per cent (platinum spiral, faint incandescence), 8.6 (spiral plunged in a hydrogen flame), 17.2 and 20.3 (hydrogen flame; the 20.3 on a later "humid day"), against 0 for dry air. That is a factor 3.5 from source and day alone, with no humidity stated for any run. Solid sources are absorbed more as they get cooler (Leslie cube at 212 F highest; p.423, 425), which favours "at least" for a 15 C ground; the highest figures need a source emitting in water-vapour bands (a hydrogen flame), which wet ground does not. His step from "water is opaque to vapour's radiation" to "vapour is opaque to water's radiation" (para 544) is an inference; no experiment with a liquid-water source is reported [read: p.434]. Also: the polished tube lengthens the effective path (deep reader's inference). A 288 K line-by-line calculation remains the deciding check.
4. Claim 2, CO2 "minor": the Historically superseded verdict stands, with Tyndall's own figures added. With a carbon-monoxide-flame source, CO2 absorbed 48.0 to 74.3 per cent (1 to 10 inches) and olefiant gas 23.2 to 65.5; the small CO2 content of laboratory air absorbed 13.8 per cent [read: image p.430-431]. For solid sources CO2 is a feeble absorber; for a CO2-matched source it is "enormous" (his word). His minor-CO2 reading is therefore true of the sources he used for the atmosphere question and not of a matched one. Whether the "dry air" reference in the hydrogen-flame runs also had CO2 removed is not stated, so a small CO2 share may sit inside 17.2 and 20.3 (deep reader's inference).
5. Claim 2, meteorological evidence: beyond Strachey, Tyndall's field evidence in Ch. XIV is other people's and carries no humidity measurement: Wells 1818 (screens and a cloud cut chilling, 12 F falling to 2 F), Glaisher's height profile, Pouillet's swansdown (7.1 and 7.21 C in April and June), Melloni's 2 C limit and "cold aerial bath", Melloni's lunar 3 to 4 degree deflection [read: Ch. XIV deep read]. His own contributions are interpretive: vapour as radiator (para 647), the dry-air supplement to Wells on Bengal ice (para 655, resting on one quoted remark of Sir Robert Barker that the best ice nights have "very little dew ... after midnight"; documented in that one published description, no hygrometer reading, and Tyndall concedes Wells's conduction reason "may have been the case"), the attribution of Pouillet's 25 and 40 per cent solar absorption to vapour (para 684, "Doubtless", his gloss), and an unfinished lunar-reflector trial (para 662). Kept in the record next to the laboratory data. These are solar figures and must not be pooled with the "ten feet" terrestrial figure. Verdict for 2A unchanged: Established, resting on the laboratory absorption data and later spectroscopy, not on this field material.
6. Claim 2, Tyndall's unsourced statements restored (qe.md retention item): "35 per cent" for a 2 inch layer of air saturated with ether vapour against planetary radiation (On Radiation p.44; Heat para 546, "I find"; no table, source or temperature; the work he answers is unnamed, probably Whewell 1853 [memory: medium]). An olefiant-gas layer 1 inch thick absorbs 26 per cent (para 545; pressure and source of the thin-layer figures not stated). The Sahara night ice (Heat p.395) has no source in the folder. Each is Tyndall's own statement with no experimental record in the folder; none is rejected. Verdict for each: Never properly tested (no independent test on record).
7. Claim 2, evidence against (qe.md retention item): Ferrel, quoted by Very, accepted Magnus's film explanation (sites.md). It is a published scientific voice on the opposing side and belongs beside Magnus, Buff and Lecher-Pernter above. Magnus beyond the first dispute: Tyndall's closing note (Heat p.533) says Magnus published a further paper "controverting" the vapour results, "by which I am not convinced", promising to examine it; the book gives none of its content and no date, so it is unknown whether it is the 1861/62 paper or a later one [read: Magnus deep read]. The only record of that paper in the folder is Tyndall's one sentence (an interested party). New Fragments p.256 adds that Magnus's radiant-heat work in gases occupied his last years "in discussion with Tyndall". Separately, Magnus held that hydrogen conducts heat like a metal and Tyndall argued convection (Heat paras 291-295); modern kinetic theory accepts gaseous conduction [memory: medium, not checked], so Tyndall's scepticism there appears to have been wrong. This bears on weighing his judgement of Magnus, not on vapour absorption. The dispute therefore does not go silent in 1868 and Tyndall's reply to Magnus's later paper was deferred, a gap rather than suppression.
8. Claim 2, Janssen (qe.md item 4): Tyndall's only record is one sentence: in summer 1866 on the Titlis he "repeated with a hand spectroscope the observations of M. Janssen on the absorption-bands of aqueous vapour" [read: New Fragments p.472]. No result, wavelength or inference is given, so it is a documented attempt, not an observation, and it is not counted as evidence. Silence about the result is not evidence that the bands were not seen. Janssen's telluric-line work [memory: medium] is a lead for the next Investigator run.
9. Claim 3, arc and solar figures: unchanged. [read: Ch. XIII deep read] Luminous fractions confirmed on images: white-hot platinum 1 in 24 (the same chapter's table prints 4.6 per cent absorbed, 1 in 21.7, not reconciled), gas flame 1 in 25, arc 1 in 10 and, after later experiments, transmission 89 and absorption 11, hence "eight times" (a 1 in 10 would give nine). The arc claim should say 8 (from 89/11) and 7.7 (area). Tyndall's alum "correction" (20 per cent of the obscure heat passes concentrated alum, p.483) corrects Melloni's alum method only; Melloni's oil flame 10/90, platinum 2/98, alcohol 1/99 therefore understate the obscure share, and the iodine and area methods are untouched. The 20 per cent is Tyndall's figure from a calibration table not in the folder. The solar "about 2" is Muller's figure measured at the ground, after atmospheric vapour, which Tyndall himself invokes to explain why it is lower than the arc's (para 573); my [calc] comparison with about 1.0 above the atmosphere therefore makes the discrepancy larger, not smaller, since vapour only removes infrared. The label stays Never properly tested. The only dated solar trial is 15 Nov 1864 (cloudy, smoky London); the others carry no dates. Ox-eye: Tyndall reports that one fifth of obscure rays reach the retina, a single report with no instrument or thickness (arithmetic gives "nearly two-thirds" of the energy reaching the retina as non-visual, checked). The qe.md suggestion of a platinum-melting-point test: a ratio of 23 corresponds to about 2400 K, above platinum's melting point near 2041 K [calc, QE; memory: high on the melting point]; the two competing readings (iodine method under-counts visible light, or platinum's emissivity is higher in the visible) are untested. Safety: Tyndall put his own eye to the dark focus (para 612a) and warned against repeating it.
10. Claim 5, weak form and water: the ordering evidence does not include water vapour [read: image p.418 footnote "Aqueous vapour, unmixed with air, condenses so readily that it cannot be directly examined"]. For ten non-water liquids and vapours, rank order agrees once the quantities are equalised (the top four vapours span 1.2 points, no errors stated), with one source ("for heat of the same quality" is Tyndall's own qualifier); the equal-quantity vapours correspond to about one-tenth of the thinnest liquid layer, so it is a rank test, not equal absorption per molecule [calc, deep reader]. Water is an extrapolation of rank (para 515: "if no single experiment on the vapour of water existed, we should be compelled to conclude ..."), plus the qualitative hydrogen-flame pair (humid air 17.2 to 20.3 per cent; liquid water 0.02 inch transmits 5.8). The source dependence of the vapour order (chloroform and iodide of methyl; formic and sulphuric ether; carbon disulphide and chloroform) is Tyndall's own evidence that a vapour's position is not a single property. Printed inconsistencies kept: formic and sulphuric ether values swapped between p.416, p.423 and p.426; hydrogen-flame water transmission 2.8 and 1.1 (p.434) against 5.2 and 2.0 (p.436) at 0.04 and 0.07 inch. The weak-form verdict (Plausible, untested independently) stands, now limited to non-water vapours.
11. Claim 5, strong form: web check done. [web: https://en.wikipedia.org/wiki/Water_absorption] states that in liquid water rotational transitions are effectively quenched ("no rotational fine structure"), the bands are broadened by hydrogen bonding, the stretch bands move from 3657 and 3756 cm-1 (gas) to 3450 and 3615 cm-1 (liquid) and the bend from 1595 to 1640 cm-1. This confirms that band shapes and positions are phase-dependent. It does not itself measure integrated broadband absorption per molecule, which is what Tyndall's weak form concerns, and Tyndall did not claim the strong form in this wording. I keep "Tested, not supported" for the strong form (band-by-band invariance across phase, including water) and the source is now [web] rather than [memory]; the wider statement that the liquid absorbs across 8 to 12 micrometres where the vapour is nearly transparent is still [memory: high], not fetched.
12. Claim 4/5 wording: qe.md asks the primer to say why "Internally inconsistent" applies to "red waves immensely more powerful than blue": it contradicts Tyndall's own wave theory, in which energy per wave scales with frequency (mathematician). Label stands.
13. Claim 6: "150 of the lower degrees of the galvanometer" for magnetised obscure heat is image-confirmed (Six Lectures p.215; the digest's "may be 15" is wrong). Still single-source (Tyndall) and the deflection scale is non-linear; the weighing in claim 6 above is unchanged.
14. Peripheral items added to the record (qe.md section 3, none changes a verdict): breath CO2, three pairs chemical against physical, 4.311/4.00, 4.66/4.56, 5.33/5.22 [read: image p.433] (differences 7.2, 2.1, 2.1 per cent; Tyndall: "very fair"; the contrarian's "within 2 per cent" applies to the last two only, and the 1865 lecture's second pair reads 4.58 rather than 4.56, a real difference between printings); Tyndall's concession of priority on singing flames to Schaffgotsch (Heat p.293), which I weigh beside, not instead of, the Akin material; the Moreno/Kirchhoff priority footnote (Six Lectures pp.288-290) and Tyndall's 1861 attempt to reverse the sodium lines (p.229); Wells's dew theory, Glaisher's table (corrected to 570 and 52), Pouillet; perfumes (17 oils, patchouli 30 to aniseed 372 times dry air, herbs 32 to 53, the thyme 74 against 68 conflict in the printed book; unmeasured quantities; Tyndall's "multiplied by millions" is an estimate); the long and short column test (olefiant retains 0.62, ethers 0.17 to 0.21 in a 3 inch tube, a stronger result than the long-tube margin of 1 to 7 degrees).

Corrected claim list: additions

• Claim 1: reciprocity is rank-confirmed and roughly proportional, not equal (item 1). Ozone numbers are preparation rankings (item 2).
• Claim 2: "neither is what he claimed" (my earlier item 4) withdrawn; his claim is a lower bound that every extrapolation exceeds (item 3). Add the 5.8 / 8.6 / 17.2 / 20.3 per cent source series. The Pouillet 25 and 40 per cent are solar figures, not evidence for the terrestrial claim (item 5).
• Claim 3: say 8 (from 89/11) and 7.7; "sun about 2" is a ground-level ratio (item 9); the alum correction concerns Melloni's method only.
• Claim 5: weak form limited to non-water vapours; water vapour is not in the tests (item 10).
• Claim 1, image checks: CO2 CO-flame table (48.0 to 74.3) and olefiant (23.2 to 65.5) confirmed; remove the "OCR, not image-checked" caveat everywhere (done in this file).
• Glaisher table: use 570 (two inches) and 52 (six feet), not the OCR 670 and 62.

Corrections to earlier documents (QE revision)

1. This file, item 4 above and claim 2B: "neither is what he claimed" withdrawn (item 3).
2. This file, source limits and claim 2E: the OCR caveat on the CO2 table removed; image-confirmed (Heat p.431).
3. digest Six-lectures part 06, p.214 ("150 may be 15"): wrong; image shows 150.
4. digest Heat part 13, p.433 and the On Radiation deep read: breath values 4.11 / "6.22" are OCR errors; correct is 4.311 / 4.00 and 5.33 / 5.22; breath pressures are 1, 3, 6 and 30 inches, not 6.5. The 4.56 (Heat) against 4.58 (On Radiation) is a genuine difference between printings.
5. physicist.md item 3 ("Heat paras 435-442, headings only"): ozone and perfumes are now read; add the numbers with the caveats in item 2.
6. claims.md and facts.json: Ch. XII, XI (paras 435-466), XIII and XIV (647-684) are no longer digest only; the Heat entry's how_read must change. Add the hydrogen-flame series, planetary shell figures, luminous fractions, alum correction, Pouillet and the Titlis mention (as a mention only, not evidence).
7. Digests that quote the Glaisher table from OCR (670, 62) should read 570 and 52; the orchestrator should grep digest/ and primer.md for these.
8. primer.md Part C Step 2 ("Neither matches Tyndall's printed claims"): same wording error as item 1. Primer should add the source-dependence series, define HITRAN and line-by-line calculation, and carry the restored Ferrel, Babbitt and 35 per cent ether-shell items (qe.md section 7).

Still open after this revision

• Humidity of any of the 4 ft laboratory runs; the dried-air reference's CO2 status.
• The date and content of Magnus's later vapour paper.
• Whether a 288 K line-by-line calculation supports "at least 10 per cent in ten feet" for typical air.
• The calibration tables behind 1 in 24, 1 in 25, 1 in 10, and the 20 per cent alum figure.
• Janssen's telluric-line work (next Investigator run) and the Whewell identification [memory: medium].

Web checks (QE revision): 1 done (Water absorption page, above). Budget otherwise unused; the Hottel and Janssen checks were not retried.